Most AMC selections happen badly, and not because anyone is careless. They happen badly because the decision gets made on a capability deck and a reference call, and the things that determine whether the relationship works are not visible in either.
This is a process you can run in about eight weeks. It is written for residential mortgage lenders — banks, credit unions and independent originators.
Before you start: decide what problem you are solving
Write down, in one paragraph, what is wrong with your current arrangement. The answer shapes everything downstream and it is frequently not what people assume.
"Turnaround is too slow" is usually several different problems wearing one coat. If the delay is concentrated in specific counties, you have a coverage problem and need a provider with depth there. If it is spread evenly, you may have an intake problem that no vendor change will fix. If it appears mostly in revisions, the issue is quality and review.
Pull your last hundred files and look before you write a single requirement.
Weeks 1–2: build a longlist
Six to ten providers is plenty. Sources worth using:
- Industry directories, for discovery only — inclusion is generally paid and ordering rarely reflects assessment
- Referrals from institutions of similar size, geography and product mix
- Your loan origination system vendor's integration partner list, which at least confirms technical compatibility
- Your existing provider, if you are testing the market rather than replacing them
At this stage you are only establishing that each provider exists, operates in your states, and handles residential volume at your scale.
Week 3: verify before you engage
Do this before any sales conversation, because it is faster than being sold to and it removes candidates cheaply.
Check each provider against the Appraisal Subcommittee national registry, which confirms federal registration. Then check the state appraiser regulatory agency register in each state where you originate, which confirms state registration and surfaces any disciplinary history.
Providers failing either check come off the list. Providers registered in most but not all of your states stay on, with a note — it may be resolvable, but it needs answering.
Weeks 4–5: the request
Send a short, specific document rather than a long generic one. Vendors respond better to ten questions requiring real answers than to sixty requiring boilerplate.
Ask for:
- Appraisers who accepted and completed an assignment in each of your top twenty counties in the last ninety days, as a count per county
- Average time from order placement to assignment acceptance, reported separately from total turnaround, for the last twelve months
- Average total turnaround for the same period, by state
- Revision rate, broken down by cause
- Evidence of current state registration in every state where you originate
- Their written revision protocol, as a document
- How they establish customary and reasonable fees, and how they evidence compliance
- Average appraiser payment terms, and actual average days to pay
- Their escalation path and named contacts, with response commitments
- Integration status with your specific loan origination system, and whether it is certified or bespoke
Question eight is the one that surprises people. Appraiser payment performance is operationally predictive: appraisers accept assignments preferentially from providers who pay promptly, so slow payers show longer acceptance times before anyone explains why. It is also the question vendors least expect from a lender.
A provider who cannot answer questions one, two and four has not examined their own operation that way. That is information.
Week 6: score it honestly
Agree the weighting before you see the responses, not after. Otherwise the weighting drifts toward whoever presented best.
A workable split for residential lending:
- Coverage in your actual footprint — 35%
- Turnaround performance, weighted toward acceptance time — 25%
- Quality and revision management — 15%
- Compliance and registration — 10%
- Technology and integration — 10%
- Cost — 5%
Cost being last is deliberate. The difference in fee between credible providers is small relative to the cost of a file closing two weeks late, and a provider winning on price frequently does so by underpaying appraisers, which converts directly into the acceptance delay you were trying to solve.
Week 7: pilot, do not switch
Never move your whole book on a presentation. Run a pilot.
Design it properly:
- Thirty to fifty orders, enough to be meaningful
- Deliberately include your difficult markets, not just your easy ones. A pilot run entirely in metropolitan counties proves nothing you needed to know
- Four to six weeks, so revision cycles have time to appear
- Measure the same metrics you asked for in the request, so you can compare claim against performance
- Run it alongside your incumbent where possible, on comparable files
The single most useful pilot metric is the gap between what was claimed in week five and what happened in week seven. A provider whose actual acceptance times match their stated averages is telling you something about how they report generally.
Week 8: contract terms worth the argument
Four provisions matter more than fee:
Service levels with definitions. A turnaround commitment is meaningless unless the clock start and stop points are defined, and unless access delays and revision cycles are explicitly carved in or out.
Reporting obligations. Specify what you receive and how often — stage-level turnaround, coverage by county, revision rate by cause. If it is not in the contract you will be asking for it as a favour.
Data portability and exit. Appraisal reports are records you may need to produce years later. Establish now what you get on termination, in what format, and within what period.
Notification of registration lapse. A clause requiring the provider to notify you if any state registration lapses or is suspended. Most contracts are silent on this, and the alternative is discovering it during an examination.
After: the part everyone skips
Set the annual review date at signing and put it in a calendar with a named owner.
Re-verify state registrations, re-run the coverage question against your current footprint — which will have shifted — and compare performance against the pilot baseline. A provider whose county coverage has quietly thinned will not raise it unprompted.
Vendor due diligence is not an onboarding event, and examiners treat it as an ongoing obligation. More practically: the failure mode in this relationship is not a bad provider chosen badly. It is a good provider whose performance drifted while nobody was measuring.
Lenders Allies provides residential appraisal management services to banks, credit unions and mortgage originators across the United States. We would encourage asking the ten questions above of any provider under consideration, ourselves included.
General information only, not legal or compliance advice.